Software Engineer (Agentic AI, LLM, Python | 5 to 8 Years)

Cisco is hiring a hybrid Software Engineer in Bangalore, India to build Agentic AI and LLM systems alongside backend, data and cloud platforms. The role focuses on scalable, user-centered software developed with secure coding practices and attention to privacy.

Responsibilities

• Lead the design of Python backend systems and data pipelines, and develop complex Agentic AI and LLM systems using LangGraph.
• Set standards for AI and RAG architectures, APIs, databases and distributed systems. Define strategies to optimize LLM tokens, latency, throughput, performance and cost.
• Drive AWS and cloud-native architecture using S3, SQS, SNS, Lambda, Docker and Kubernetes. Resolve development and production issues across teams.
• Work with design, product management and other stakeholders to align technical solutions with customer needs and business goals. Produce technical and user documentation, mentor engineers and improve engineering practices.

Minimum qualifications:

• The posting specifies a minimum of 5 to 10 years of hands-on experience in machine learning engineering, backend development and applied AI; the job title says “5 to 8 Years.”
• Deep experience with Agentic AI, LLM applications, LangGraph and RAG; production deployment of deep learning and LLM models; and deployment of LangGraph and multi-agent workloads with versioned prompts, runtime safeguards and rollback strategies.
• Strong Python, backend engineering and distributed systems expertise; LLM optimization and evaluation or observability tooling such as LangSmith; and ETL optimization techniques including workload-aware scheduling, partition pruning, caching and storage-format tuning.

Preferred qualifications include mentoring and leading cross-functional initiatives, aligning technical strategy with stakeholder priorities, translating complex requirements into scalable solutions, knowledge of networking and enterprise integration patterns, strong written and verbal communication, and experience integrating Model Context Protocol and related AI tools. This is a full-time role listed at 40 hours per week.